MDS in sklearn

[stevenxxiu]

First set of features to try:

Location (coordinates and euclidean distance)
Population Density:             
Travel time to GPO (minutes)
Distance to GPO (km)
LGA
Primary Care Partnership
Medicare Local
Area (km^2)
ARIA+ (min)
ARIA+ (max)
ARIA+ (avg)
ABS remoteness category
DHS Area
Commercial (km^2)
Commercial (%)
Industrial (km^2)
Industrial (%)
Residential (km^2)
Residential (%)
Rural (km^2)
Rural (%)
Other (km^2)
Other (%)

We'll use l2 distance at first and, and then try:

• weighting features
• removing redundant features (equivalent to a weight of 0)
• other distances metrics

[dsamores]

MDS with Euclidean distance is done now. I'm plotting the resulting data. Please check, you can run the code yourselves or take a look at the image attached. It's hard to distinguish some labels in that image, but if you run the code you can zoom in the chart.

I'm not really sure if the results make sense or if I'm doing or understanding something wrong. Please let me know your thoughts.

Labels are outputted when the program is run, but I'm attaching a text file with them as well.

Note: Steven if I'm doing something weird in Python (like the camel case thing) please let me know.

labels.txt